Joint Features Extraction for Multiple Moving Targets Using (Ultra-)Wideband FMCW Signals in the Presence of Doppler Ambiguity

This article addresses the joint estimation of range, velocity and azimuth for multiple fast-moving targets using (ultra-)wideband (UWB) frequency-modulated continuous-wave (FMCW) radar with a phased array in the presence of Doppler ambiguities. The range migration of moving targets is described with the coupling of the fast-time and slow-time (chirp index), leading to the smearing of the target Doppler spectrum. This phenomenon degrades the performance of conventional detection and estimation techniques. As with range-Doppler processing, the estimation accuracy for direction-of-arrival (DOA) with conventional narrowband-based algorithms significantly degrades if a UWB signal is deployed. For the FMCW waveform, the wideband DOA differs from the narrowband DOA due to an extra coupling term, similar to the range migration problem. A novel spectral norm-based algorithm for joint estimation of range, velocity and DOA of fast-moving targets is proposed, taking the appropriate wideband signal model with the coupling terms into account. The proposed spectral norm-based algorithm avoids off-grid peak search and can be easily accelerated with power iteration; it outperforms conventional coherent integration methods in both accuracy and efficiency when using moderate data size. The advantages of the proposed algorithm and its super-resolution capability are validated with numerical simulations.